Re: [RFC][BUGFIX] memcg: rmdir doesn't return

[KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

On Mon, 15 Jun 2009 11:50:21 +0900
Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> wrote:


> > Then, my suggestion is here.
> > ==
> > } else {
> > 	ent.val = page_private(page);
> > 	id = lookup_swap_cgroup(ent);
> > 	rcu_read_lock();
> > 	mem = mem_cgroup_lookup(id);
> > 	if (mem) {
> > 		if (css_tryget(mem->css)) {
> > 			/*
> > 			 * If no processes in this cgroup, accounting back to
> > 			 * this cgroup seems silly and prevents RMDIR.
> > 			 */
> > 			struct cgroup *cg = mem->css.cgroup;
> > 			if (!atomic_read(&cg->count) && list_empty(&cg->children)) {
> > 				css_put(&mem->css);
> > 				mem = NULL;
> > 			}
> > 	}
> > 	rcu_read_unlock();
> >  }
> > ==
> > 
> Thank you for your suggestion.
> To be honest, I think swap cache behavior would be complicated anyway :(
> 
> I prefer my change because the behavior would become consistent with
> the case we don't use mem+swap controller and with the behavior of page cache.
> 
I don't like implict resource move. I'll try some today. plz see it.
_But_ this case just happens when swap is shared between cgroups and _very_ heavy
swap-in continues very long. I don't think this is a fatal and BUG.

But ok, maybe wake-up path is not enough.

Thanks,
-Kame

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>